After many years and product marketing campaigns, defining a concept like artificial intelligence without going too broad or narrow is still tricky business for both governing bodies and consumers alike. But regulators looking to mitigate the risks AI can pose in areas like education, law enforcement or even employment decisions, all without stymying technological innovation, may find themselves consistently erring on the side of “broad.”

This may already be the case with the initial proposal for the Artificial Intelligence Act (AIA) unveiled by the European Commission last April. The proposed legislation attempts to thread the needle between a broad and narrow overview by defining AI according to four different risk categories: unacceptable risk, high risk, limited risk or minimal risk.